Ontology Alignment-based Method Of Constructing Electronic Medical Records

With the deepening of medical informatization，many hospitals andorganizations have their own information system，but still can’t exchangedata with each other. This leads to medical data has big capacity but lowutilization. In consider of this issue, we choose ontology——the sharedconcept model as the presentation of medical data, propose an ontologyalignment-based method of constructing electronic medical records. Atfirst, original data should be transformed to Linked-Data, and then initialontology can be extracted from Linked-Data. At last, we align initialontologies to EHR.In this paper, the following tasks will be completed.1) The Linked-Data mapping rule of medical relational data.We analyze the medical database, as well as the existed medical domain,replenish the Linked-Data mapping rule for medical relational data.2) The Instance-based construct method of ontologyWe cluster instances with breadth-first cluster method, extract conceptfrom clusters with the principle that concept is a collection of instances, identify the superclass relationship of concepts with the supersetrelationship of instance set, and build ontology concept hierarchy.3) Instance similarity-based alignment method of ontologyThe traditional ontology alignment is extracting logical informationbased on abstract concepts. Our approach is just based on the fact, guidedby the instance data, and mine concept semantic from instance. And wepropose the concept of property semantic, replenish the concept semanticbetween instance level and concept level with property semantic, optimizethe concept semantic and semantic query.An EHR system is implemented to verify the validity of the methodusing actual medical data. This system proved that the method canintegrate medical data effectively and provide a satisfied semantic queryresult.